Note: this INT8 version of Llama 3.2 3B has been quantized for inference on Intel CPU, Intel iGPU (Arc) via intel-extension-for-pytorch, Intel NPU (AI Boost on Core Ultra series) via OpenVINO.

Usage

This is a base / completion model — prompt it directly with raw text:

python

from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "fbaldassarri/meta-llama_Llama-3.2-3B-auto_gptq-int8-gs64-asym"
model = AutoModelForCausalLM.from_pretrained(repo, device_map="auto")
tokenizer = AutoTokenizer.from_pretrained(repo)

prompt = "The quick brown fox"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=64)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Replication Recipe

The recommended way to reproduce this exact quantization is via the auto-round-pipeline — the same orchestration tool that produced this artifact.

Step 1 — Bootstrap the auto-round-pipeline

Set up a dedicated conda environment using the pipeline's setup.sh. Any of the install modes below produces an environment that can reproduce this quantization; pick the one that matches your goals:

git clone https://git.epicdynamic.com/auto-round-pipeline
cd auto-round-pipeline

# Pinned PyPI wheel (fastest; matches what this pipeline used by default):
bash setup.sh --pip-version 0.13.1

# Or build from intel/auto-round at the same tag (byte-identical reproducibility):
bash setup.sh --source-tag v0.13.1

# Intel Arc iGPU acceleration (e.g. Core Ultra 185H) — append to either of the above:
#   ... --intel-xpu
# NVIDIA / AMD opt-in: --cuda / --rocm

The script prints the resulting conda env name (something like auto-round-pipeline-v0.13.1[-src][-xpu|-cuda|-rocm]) at the end.

Step 2 — Quantize just this model

Activate the env that setup.sh created, then invoke the runner with the same job filters that produced this artifact:

conda activate <env-name-printed-by-setup.sh>
python runner.py \
    --model 'meta-llama/Llama-3.2-3B' \
    --quant 'INT8-gs64' \
    --format auto_gptq \
    --no-upload      # drop this to also push to HuggingFace Hub

Step 3 — (Optional) standalone Python recipe

If you'd rather call auto-round directly without the orchestration wrapper, this is the exact call the pipeline made:

python

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from auto_round import AutoRound

model_name = "meta-llama/Llama-3.2-3B"
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16)
tokenizer = AutoTokenizer.from_pretrained(model_name)

bits, group_size, sym = 8, 64, False
autoround = AutoRound(
    model, tokenizer,
    bits=bits, group_size=group_size, sym=sym,
    device_map="cpu",
    nsamples=128, iters=200, seqlen=512, batch_size=4,
)
autoround.quantize_and_save("./AutoRound/meta-llama_Llama-3.2-3B-auto_gptq-int8-gs64-asym", format="auto_gptq")

Actual Run Conditions

Recorded by the auto-round-pipeline at quantization time:

Field	Value
Intel auto-round version	0.13.1
transformers version	4.55.3
torch version	2.12.1+cpu
torch_dtype (load)	torch.bfloat16
calibration device	`cpu`
calibration samples	128
tuning iterations	200
calibration seq len	512
calibration batch size	4
quantization duration	20672.7s (344.5 min)
completed at (UTC)	2026-07-10T20:33:32.152281+00:00

License

Llama 3.2 Community License

Disclaimer

This quantized model comes with no warranty. It has been developed only for research purposes.